Image Processing Reference
In-Depth Information
u
a
Cp
u
a
,
B
Cp
u
j
,
B
U
1
¼
2
max
j¼
1
...
k
1
ð
12
:
8
Þ
• Step 2: Based on Eq. (
12.7
), for each user
u
a
of
U
1
, select a group
U
2
with the
k
2
users that have the strongest and most often interaction with user
u
a
and their
connectivity variation pattern shows that they are most often connected from a
given location.
u
b
LDI
p
u
a
,
u
b
LDI
p
u
a
,
u
j
U
2
¼
2
max
ð
Þ
U
1
,
j¼
1
...
k
2
u
a
2
• Step 3: Based on Eq. (
12.4
), for each user
u
b
of
U
2
, select a set
X
n
with the
n
nodes that most often user
u
b
connects to the framework using this node as
access node.
x
i
Cp
u
b
,
x
i
u
b
2U
2
,
j¼
1
...
n
Cp
u
b
,
x
j
X
n
¼
2
max
ð
Þ
ð
12
:
9
Þ
• Step 4: If we define NS
p
A
,
B
the
Node Selection probability
for selecting the node
A from a node B as an optimal location for an overlay node taking into account
only the network criteria, we may select node
A
from node
B
based on the social
interaction of the users in case:
φ
1
NS
p
A
,
B
þ φ
2
NS
p
A
,
x
i
>
SNS
ð
12
:
10
Þ
where SNS is a threshold,
˕
2
are weighting factors to give emphasis to the
social dimension of the node
B
and
x
i
, and
˕
1
and
˕
1
þ ˕
2
¼
1
ð
12
:
11
Þ
Equation (
12.10
) may be also used for the overall network overlay optimization.
12.6 Validation
In this section, we provide an initial validation of the method that has been
presented. For the evaluation of the interaction probability, we use three services:
chat, talk and video conference with weights w
c
,w
t
and w
v
respectively and realize
the following scenarios (see Table
12.1
):
In all scenarios, we give more value to the video conference metric, as people
that are used to communicate via video conference would be at least the early
adopters of the 3D immersion communications system. Users that mainly chat and
not use voice or video conference today may not utilize it in the near future.
